Re: [Issue 1302] Changed - cvs2svn fails on file added on branch

Robert Pluim <rpluim@bigfoot.com> writes:
> OK, cvs2svn didn't fail this time, I just ran out of disk space :-(
> I've made some space, and started it again. (41000 revs and counting)
>
> This is going to be a real problem for people with large repos, since
> the dump-file contains a full copy of every revision of every file,
> which is going to be huge.
>
> Perhaps cvs2svn should be checkpointable in some way? Would it be
> enough to split the cvs2svn-data.*revs files at suitable places and
> run pass 4 & 5 multiple times?

Hmmm. I think I'd rather offer an option to split the dumpfile, which
is an entity people may know how to work with anyway. You give it N,
it writes out dumpfiles of N revisions at a time, on successive
invocations.

(But, that's not to be tackled until after all functionality is in,
IMHO.)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed May 21 20:31:33 2003